You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
isc: two-sample permutation test fails for pairwise ISCs #449
When computing pairwise ISCs, the output is a condensed triangle vector (via squareform) with length N * (N - 1) / 2, where N is the number of subjects. In the context of two separate (non-paired) groups, ISCs are computed separately and the resulting condensed vectors are v-stacked (note that there may not be the same number of subjects in each group). When passing this stacked pairwise ISC array for two groups to permutation_isc, the _check_isc_input doesn't handle the stacked groups properly. It assumes there's one group and breaks because that stacked array likely won't have the correct number of elements for squareform. I think most of the rest of the permutation test handles this situation properly, but the logic of _check_isc_input needs to be adjusted, and we need a test confirms this is handled properly. Thanks to @mcdoar9 for catching this!
We've confirmed that this is not really a bug; more that the documentation is unclear as to how two-sample pairwise ISCs are to be supplied to permutation_isc. PR #476 should clarify this, so I'm closing this.
Activity
snastase commentedon Jul 31, 2020
We've confirmed that this is not really a bug; more that the documentation is unclear as to how two-sample pairwise ISCs are to be supplied to
permutation_isc. PR #476 should clarify this, so I'm closing this.